Image registration device, method, and program

1. An image registration device, comprising: a storage device storing an image registration program; and a central processing unit (CPU) configured to execute the image registration program to: acquire a plurality of images captured in time series, the plurality of images comprising a first image including a first a plurality of pixels located at first pixel positions in the first image, a second image including a second plurality of pixels located at second pixel positions in the second image, and a third image including a third plurality of pixels located at third pixel positions in the third image, the first pixel positions corresponding to the second pixel positions and to the third pixel positions; acquire pixel value changes between the first plurality of pixels located at the first pixel positions, the second plurality of pixels located at the second pixel positions, and the third plurality of pixels located at the third pixel positions; cluster a first pixel value change of the pixel value changes in to a first class, the first pixel value change representing pixels of the first plurality of pixels, pixels of the second plurality of pixels, and pixels of the third plurality of pixels having a same first pixel value change, the first class corresponding to a right atrium of a heart and a right ventricle of the heart; cluster a second pixel value change of the pixel value changes in to a second class, the second pixel value change representing pixels of the first plurality of pixels, pixels of the second plurality of pixels, and pixels of the third plurality of pixels having a same second pixel value change, the second class corresponding to a left atrium of the heart and a left ventricle of the heart; cluster a third pixel value change of the pixel value changes in to a third class, the third pixel value change representing pixels of the first plurality of pixels, pixels of the second plurality of pixels, and pixels of the third plurality of pixels having a same third pixel value change, the third class corresponding to a myocardium of the heart; divide each image of the plurality of images into a plurality of regions corresponding to the first class, the second class, and the third class, the dividing based on the clustering the first pixel value change, the clustering second pixel change, and the clustering the third pixel value change; and perform registration processing on each image of the plurality of images based on the dividing each image of the plurality of images into the plurality of regions.

2. The image registration device according to claim 1 , wherein, in a case where a statistical value of a pixel value associated with the first class is equal to or less than a threshold value set in advance, the CPU replaces pixel values associated with the first class with pixel values associated with the third class, and divides a region corresponding to the first class and a region corresponding to the third class as a same region.

3. The image registration device according to claim 1 , wherein, in a case where a statistical value of a pixel value associated with the second class is equal to or less than a threshold value set in advance, the CPU replaces pixel values associated with the second class with pixel values associated with the third class, and divides a region corresponding to the second class and a region corresponding to the third class as a same region.

4. The image registration device according to claim 1 , wherein the CPU divides each image of the plurality of images into the plurality of regions by performing graph cut.

5. The image registration device according to claim 1 , wherein the CPU replaces a pixel value associated with each region of the plurality of regions with a representative value, the representative value being different for each region, and the performed registration processing is based on the replaced pixel value associated with each region.

6. The image registration device according to claim 1 , wherein the CPU performs the clustering based on at least one of a peak value or a time at which the peak value is obtained.

7. An image registration method, comprising: acquiring a plurality of images captured in time series, the plurality of images comprising a first image including a first a plurality of pixels located at first pixel positions in the first image, a second image including a second plurality of pixels located at second pixel positions in the second image, and a third image including a third plurality of pixels located at third pixel positions in the third image, the first pixel positions corresponding to the second pixel positions and to the third pixel positions; acquiring pixel value changes between the first plurality of pixels located at the first pixel positions, the second plurality of pixels located at the second pixel positions, and the third plurality of pixels located at the third pixel positions; clustering a first pixel value change of the pixel value changes in to a first class, the first pixel value change representing pixels of the first plurality of pixels, pixels of the second plurality of pixels, and pixels of the third plurality of pixels having a same first pixel value change, the first class corresponding to a right atrium of a heart and a right ventricle of the heart; clustering a second pixel value change of the pixel value changes in to a second class, the second pixel value change representing pixels of the first plurality of pixels, pixels of the second plurality of pixels, and pixels of the third plurality of pixels having a same second pixel value change, the second class corresponding to a left atrium of the heart and a left ventricle of the heart; clustering a third pixel value change of the pixel value changes in to a third class, the third pixel value change representing pixels of the first plurality of pixels, pixels of the second plurality of pixels, and pixels of the third plurality of pixels having a same third pixel value change, the third class corresponding to a myocardium of the heart; dividing each image of the plurality of images into a plurality of regions corresponding to the first class, the second class, and the third class, the dividing based on the clustering the first pixel value change, the clustering second pixel change, and the clustering the third pixel value change; and performing registration processing on each image of the plurality of images based on the dividing each image of the plurality of images into the plurality of regions.

8. A non-transitory computer-readable recording medium having stored therein an image registration program causing a computer to: acquire a plurality of images captured in time series, the plurality of images comprising a first image including a first a plurality of pixels located at first pixel positions in the first image, a second image including a second plurality of pixels located at second pixel positions in the second image, and a third image including a third plurality of pixels located at third pixel positions in the third image, the first pixel positions corresponding to the second pixel positions and to the third pixel positions; acquire pixel value changes between the first plurality of pixels located at the first pixel positions, the second plurality of pixels located at the second pixel positions, and the third plurality of pixels located at the third pixel positions; cluster a first pixel value change of the pixel value changes in to a first class, the first pixel value change representing pixels of the first plurality of pixels, pixels of the second plurality of pixels, and pixels of the third plurality of pixels having a same first pixel value change, the first class corresponding to a right atrium of a heart and a right ventricle of the heart; cluster a second pixel value change of the pixel value changes in to a second class, the second pixel value change representing pixels of the first plurality of pixels, pixels of the second plurality of pixels, and pixels of the third plurality of pixels having a same second pixel value change, the second class corresponding to a left atrium of the heart and a left ventricle of the heart; cluster a third pixel value change of the pixel value changes in to a third class, the third pixel value change representing pixels of the first plurality of pixels, pixels of the second plurality of pixels, and pixels of the third plurality of pixels having a same third pixel value change, the third class corresponding to a myocardium of the heart; divide each image of the plurality of images into a plurality of regions corresponding to the first class, the second class, and the third class, the dividing based on the clustering the first pixel value change, the clustering second pixel change, and the clustering the third pixel value change; and perform registration processing on each image of the plurality of images based on the dividing each image of the plurality of images into the plurality of regions.

9. An image registration device, comprising: a storage device storing an image registration program; and a central processing unit (CPU) configured to execute the image registration program to: acquire a plurality of images captured in time series, the plurality of images comprising a first image including a first a plurality of pixels located at first pixel positions in the first image, a second image including a second plurality of pixels located at second pixel positions in the second image, and a third image including a third plurality of pixels located at third pixel positions in the third image, the first pixel positions corresponding to the second pixel positions and to the third pixel positions; acquire pixel value changes between the first plurality of pixels located at the first pixel positions, the second plurality of pixels located at the second pixel positions, and the third plurality of pixels located at the third pixel positions; cluster a first pixel value change of the pixel value changes in to a first class, the first pixel value change representing pixels of the first plurality of pixels, pixels of the second plurality of pixels, and pixels of the third plurality of pixels having a same first pixel value change, the first class corresponding to a right atrium of a heart and a right ventricle of the heart; cluster a second pixel value change of the pixel value changes in to a second class, the second pixel value change representing pixels of the first plurality of pixels, pixels of the second plurality of pixels, and pixels of the third plurality of pixels having a same second pixel value change, the second class corresponding to a left atrium of the heart and a left ventricle of the heart; cluster a third pixel value change of the pixel value changes in to a third class, the third pixel value change representing pixels of the first plurality of pixels, pixels of the second plurality of pixels, and pixels of the third plurality of pixels having a same third pixel value change, the third class corresponding to a myocardium of the heart; divide each image of the plurality of images into a plurality of regions corresponding to the first class, the second class, and the third class, the dividing based on the clustering the first pixel value change, the clustering second pixel change, and the clustering the third pixel value change; and perform registration processing on each image of the plurality of images based on the dividing each image of the plurality of images into the plurality of regions, wherein the CPU replaces a pixel value associated with each region of the plurality of regions with a representative value, the representative value being different for each region, and the performed registration processing is based on the replaced pixel value associated with each region, wherein, in a case where a statistical value of a pixel value associated with the first class is equal to or less than a threshold value set in advance, the CPU replaces pixel values associated with the first class with pixel values associated with the third class, and divides a region corresponding to the first class and a region corresponding to the third class as a same region, and in a case where a region corresponding to the first class in one image to be subjected to registration processing is divided as a same region as a region corresponding to the third class and a region corresponding to the first class, and a region corresponding to the third class in another image subjected to registration processing are divided as separate regions, the CPU replaces a pixel value associated with a region corresponding to the first class in the another image with a representative value associated with a region corresponding to the third class.

10. An image registration method, comprising: acquiring a plurality of images captured in time series, the plurality of images comprising a first image including a first a plurality of pixels located at first pixel positions in the first image, a second image including a second plurality of pixels located at second pixel positions in the second image, and a third image including a third plurality of pixels located at third pixel positions in the third image, the first pixel positions corresponding to the second pixel positions and to the third pixel positions; acquiring pixel value changes between the first plurality of pixels located at the first pixel positions, the second plurality of pixels located at the second pixel positions, and the third plurality of pixels located at the third pixel positions; clustering a first pixel value change of the pixel value changes in to a first class, the first pixel value change representing pixels of the first plurality of pixels, pixels of the second plurality of pixels, and pixels of the third plurality of pixels having a same first pixel value change, the first class corresponding to a right atrium of a heart and a right ventricle of the heart; clustering a second pixel value change of the pixel value changes in to a second class, the second pixel value change representing pixels of the first plurality of pixels, pixels of the second plurality of pixels, and pixels of the third plurality of pixels having a same second pixel value change, the second class corresponding to a left atrium of the heart and a left ventricle of the heart; clustering a third pixel value change of the pixel value changes in to a third class, the third pixel value change representing pixels of the first plurality of pixels, pixels of the second plurality of pixels, and pixels of the third plurality of pixels having a same third pixel value change, the third class corresponding to a myocardium of the heart; dividing each image of the plurality of images into a plurality of regions corresponding to the first class, the second class, and the third class, the dividing based on the clustering the first pixel value change, the clustering second pixel change, and the clustering the third pixel value change; and performing registration processing on each image of the plurality of images based on the dividing each image of the plurality of images into the plurality of regions, wherein the CPU replaces a pixel value associated with each region of the plurality of regions with a representative value, the representative value being different for each region, and the performed registration processing is based on the replaced pixel value associated with each region, wherein, in a case where a statistical value of a pixel value associated with the first class is equal to or less than a threshold value set in advance, pixel values associated with the first class are replaced with pixel values associated with the third class, and a region corresponding to the first class and a region corresponding to the third class are divided as the same region, and in a case where a region corresponding to the first class in one image to be subjected to registration processing is divided as a same region as a region corresponding to the third class and a region corresponding to the first class and a region corresponding to the third class in another image subjected to registration processing are divided as separate regions, the pixel value associated with a region corresponding to the first class in the another image subjected to registration processing is replaced with a representative value associated with a region corresponding to the third class.

11. An image registration device, comprising: a storage device storing an image registration program; and a central processing unit (CPU) configured to execute the image registration program to: acquire a plurality of images captured in time series, the plurality of images comprising a first image including a first a plurality of pixels located at first pixel positions in the first image, a second image including a second plurality of pixels located at second pixel positions in the second image, and a third image including a third plurality of pixels located at third pixel positions in the third image, the first pixel positions corresponding to the second pixel positions and to the third pixel positions; acquire pixel value changes between the first plurality of pixels located at the first pixel positions, the second plurality of pixels located at the second pixel positions, and the third plurality of pixels located at the third pixel positions; cluster a first pixel value change of the pixel value changes in to a first class, the first pixel value change representing pixels of the first plurality of pixels, pixels of the second plurality of pixels, and pixels of the third plurality of pixels having a same first pixel value change, the first class corresponding to a right atrium of a heart and a right ventricle of the heart; cluster a second pixel value change of the pixel value changes in to a second class, the second pixel value change representing pixels of the first plurality of pixels, pixels of the second plurality of pixels, and pixels of the third plurality of pixels having a same second pixel value change, the second class corresponding to a left atrium of the heart and a left ventricle of the heart; cluster a third pixel value change of the pixel value changes in to a third class, the third pixel value change representing pixels of the first plurality of pixels, pixels of the second plurality of pixels, and pixels of the third plurality of pixels having a same third pixel value change, the third class corresponding to a myocardium of the heart; divide each image of the plurality of images into a plurality of regions corresponding to the first class, the second class, and the third class, the dividing based on the clustering the first pixel value change, the clustering second pixel change, and the clustering the third pixel value change; and perform registration processing on each image of the plurality of images based on the dividing each image of the plurality of images into the plurality of regions, wherein the CPU replaces a pixel value associated with each region of the plurality of regions with a representative value, the representative value being different for each region, and the performed registration processing is based on the replaced pixel value associated with each region, wherein, in a case where a statistical value of a pixel value associated with the second class is equal to or less than a threshold value set in advance, the CPU replaces pixel values associated with the second class with pixel values associated with the third class, and divides a region corresponding to the second class and a region corresponding to the third class as a same region, and in a case where a region corresponding to the second class in one image to be subjected to registration processing is divided as a same region as a region corresponding to the third class and a region corresponding to the second class and a region corresponding to the third class in another image subjected to registration processing are divided as separate regions, the CPU replaces a pixel value associated with a region corresponding to the second class in the another image with a representative value associated with a region corresponding to the third class.

12. An image registration method, comprising: acquiring a plurality of images captured in time series, the plurality of images comprising a first image including a first a plurality of pixels located at first pixel positions in the first image, a second image including a second plurality of pixels located at second pixel positions in the second image, and a third image including a third plurality of pixels located at third pixel positions in the third image, the first pixel positions corresponding to the second pixel positions and to the third pixel positions; acquiring pixel value changes between the first plurality of pixels located at the first pixel positions, the second plurality of pixels located at the second pixel positions, and the third plurality of pixels located at the third pixel positions; clustering a first pixel value change of the pixel value changes in to a first class, the first pixel value change representing pixels of the first plurality of pixels, pixels of the second plurality of pixels, and pixels of the third plurality of pixels having a same first pixel value change, the first class corresponding to a right atrium of a heart and a right ventricle of the heart; clustering a second pixel value change of the pixel value changes in to a second class, the second pixel value change representing pixels of the first plurality of pixels, pixels of the second plurality of pixels, and pixels of the third plurality of pixels having a same second pixel value change, the second class corresponding to a left atrium of the heart and a left ventricle of the heart; clustering a third pixel value change of the pixel value changes in to a third class, the third pixel value change representing pixels of the first plurality of pixels, pixels of the second plurality of pixels, and pixels of the third plurality of pixels having a same third pixel value change, the third class corresponding to a myocardium of the heart; dividing each image of the plurality of images into a plurality of regions corresponding to the first class, the second class, and the third class, the dividing based on the clustering the first pixel value change, the clustering second pixel change, and the clustering the third pixel value change; and performing registration processing on each image of the plurality of images based on the dividing each image of the plurality of images into the plurality of regions, wherein the CPU replaces a pixel value associated with each region of the plurality of regions with a representative value, the representative value being different for each region, and the performed registration processing is based on the replaced pixel value associated with each region, wherein, in a case where a statistical value of a pixel value associated with the second class is equal to or less than a threshold value set in advance, pixel values associated with the second class are replaced with pixel values associated with the third class, and a region corresponding to the second class and a region corresponding to the third class are divided as the same region, and in a case where a region corresponding to the second class in one image to be subjected to registration processing is divided as a same region as a region corresponding to the third class and a region corresponding to the second class and a region corresponding to the third class in another image subjected to registration processing are divided as separate regions, a pixel value associated with a region corresponding to the second class in the another image is replaced with a representative value associated with a region corresponding to the third class.